B. Results

VI. Conclusion

The main finding is that the most important drivers of the scale of investment as measured by employment levels in firms are establishment-specific characteristics, vertical factors pertaining to local labor supply and wages, local industrial specialization, business attraction subsidies, market size, and investor country characteristics pertaining to manufacturing and trade propensity. Interestingly the measures accounting for business climate, human capital, and local information-based assets were not as significant nor robust across various

specifications nor fully in line with theoretical expectations. While more research is needed, these results suggest that the effect of business climate and human capital development on immediate job creation and the size of investment may be somewhat overstated. While these measures may play a role in the long-term growth of regional economies and overall capacity for growth or even in the global context, there does not appear to be strong statistical support for a direct role in foreign-owned firms in the U.S. over the last ten years.

There are a number of limitations to the model and results. First, the results have

somewhat narrow applications since the study’s universe of establishments is limited to firms that were at one point foreign-owned. This is a major censoring and truncation issue that limits the generalizability and scope of this study’s findings. Ultimately, this limitation makes the analysis

more exploratory in nature than a pure empirical test of the location patterns of foreign-owned firms relative to their domestic counterparts. Additional data sampling non-foreign-owned domestic establishments would enable the study to speak more generally to the underlying drivers of employment growth in firms and whether location decisions are any different for domestically-owned firms as they are for foreign-owned firms. In addition, a broader sampling universe would better isolate the effect of foreign-ownership, especially in the case of mergers and acquisitions and could reveal how domestic firms respond to foreign takeovers and if there are any systematic similarities between firms that are eventually acquired. The inclusion of an identification criterion to control for multinational corporate owners and scale economies would also help to assess the differences in investment patterns across establishments.

Another weakness is in the finding pertaining to business attraction subsidies in that it fails to take into account the selection bias within the universe of establishments that receive subsidies and the counterfactual about what would have happened had the firm not received the subsidy in the first place. A further limitation due to the exclusive focus on within-host differences is that the role of covariates that are relatively less variable in the domestic context could be understated. This is particularly true for human capital and business climate, which vary more significantly on a country to country basis, and should by extension play a larger role in the global decision making of foreign investors.

Despite these drawbacks, the advantage in evaluating employment intensity in the limited context of firms that were at one point foreign-owned is that the location decisions of foreign multinationals may reflect a site selection calculation based more on due diligence and economic fundamentals than hometown preference. As a result, in surveying the location determinants in this study one can better understand what matters to sophisticated and objective international investors and could lead to broader insights into the drivers of investment and job creation more generally.

Future research should continue to investigate the questions raised in this study. There are a number of additional covariates that were not tested and that could yield further insight including data on physical infrastructure, supply chain networks, improved source country data, various interaction terms, and foreign parent characteristics that distinguish vertical from horizontal foreign investment. For instance, future work could create an index of industrial similarity between the NAICS code of parents and subsidiaries in order to control for the relative

“horizontalness” or “verticalness” of the investment. Another improvement would be to further test the effects of information-based assets using covariates from the OECD REGPAT database which reports patenting microdata classified by companies, technologies, and metropolitan areas

or the Leiden University Centre for Science and Technology bibliographic citations database containing the number of academic publications that feature collaborations between university and private sector researchers.

Given the research potential provided by the NETS database, a number of related research questions should be explored in future work. More research is needed to understand what happens to firms following an M&A event and whether FDI is a net-positive source of job creation. A Kaplan-Meir survival analysis with cumulative hazard ratios are potential ways to evaluate the probability of a closure or layoff event following a merger. General summary statistics aggregating year-to-year changes across establishments over time may also help quantify cumulative net job impacts of FDI.

With this data it is also possible to test the probability that a firm becomes an exporter after being acquired by a foreign investor using a logistic regression approach. And given the ability to track establishment relocations it is also possible to evaluate the effect of different states on firms that changed locations. Finally, future research should continue to investigate FDI at lower levels of aggregation using covariates at the block and sub-metropolitan scale, especially to test variation in utilities costs or measures of internet connectivity and penetration.

While it is still too early to prescribe urgent policy actions, the most significant implication from this exploratory research is that a “back to fundamentals” approach that leverages workforce skills, industrial specializations, and local market capacity may represent the most empirically-grounded strategy to attract and foster growth in existing foreign-owned establishments. Although a large portion of the variation is unexplained in the econometric model, the theoretically

inconsistent or weak effects of business climate, information-based assets, and human capital suggest that more research is needed in understanding how these conditions attract international investment and their role in driving the size and growth in establishments receiving foreign investment. Again, due to the exploratory nature of the study further work is needed to assess the robustness and reproducibility of these results. At a minimum, this research should help inform future work as it offers as an exhaustive look into possible location determinants and the strength of their relationship with the employment in foreign-owned establishments over the past ten years.

In document Within-Host Location Determinants of Employment in Foreign-Owned Establishments in the U.S., 2000-2011: A survey of business climate, vertical, horizontal, and export platform motivations (Page 40-43)